**UPDATE**: tagged as easy because the new cluster should be able to do the monthly aggregation in AQS (people are already requesting this data and aggregating it themselves so it doesn't increase the load on Cassandra, just the CPU usage on AQS, for which there is plenty of overhead).
Over many years  quite a few people asked for monthly page view stats that can be queried via an API.
Since medium 2015 we have the awesome api which serves highly granular pageview data. But for trend analysis or merely monthly reports a higher aggregation level would be very useful. And in particular monthly totals.
As it turns out the current api has 'scaling issues'  .
Daily and monthly aggregates are already available, as huge downloads from https://dumps.wikimedia.org/other/pagecounts-ez/merged/
A KISS stopgap solution might be to feed those monthly dumps into a simple (SQL?) database.
keys: project, language, month, site, title, count
Capacity needed per month:
pagecounts-2016-06.bz2 contains 175 million lines and contains all titles requested, for all projects
pagecounts-2016-06-views-ge-5-totals.bz2 with only titles that were requested 5 times or more in a full month reduces this to 47 million lines. These files go back to Dec 2011. (but of course data before May 2015 were not filtered for bots)